AstraZeneca is exploring how big data and new technologies can drive innovation. They have established technology incubation labs in San Francisco, Cambridge, and Shanghai to tap into emerging technologies. The company has implemented an enterprise search platform powered by Sinequa to search over 180 million documents. This search engine underlies applications like an approvals mobile app and tools for finding experts within the company. AstraZeneca continues to enhance usability and test new approaches like predictive modeling and in-video search capabilities.
How Big Data can drive innovative technologies and new approaches in large organisations
1. How big data can drive innovative technologies
and new approaches in large organisations.
Nick Brown
Big Data In Paris, 2016 08 March 2016
2. 2
AstraZeneca
AstraZeneca is a biopharmaceutical company with R&D at its core. Our business is
providing innovative, effective medicines that make a real difference to patients.
We have grown from agrochemicals and paints, to pharmaceuticals and biologics.
But as we virtualise our activities, working increasingly with external researchers, how
we access information, integrate data and leverage knowledge is key to our success.
3. CTO Office
3
Our team was established to create new value by catalysing innovative,
emerging technologies across AstraZeneca
Technology
Incubation Labs
Competency
Centers
Enterprise
Architecture
Multi-disciplinary teams that test
new technologies and accelerate
platforms to build internal
expertise and hands-on experience
of potential game-changing
technologies whilst focusing on
immediate business problems
Established User Experience and
Mobility competency centers as
key strategic areas for future
success. They will develop and be
embedded into enterprise
capabilities with world-class
technology leadership
Providing enterprise leadership to
understand the pain points of the
business and ensure that proposed
technology changes are unified and
governed to maximise business
value as AZIT develop corporate
platforms, workflows and choices
4. Global Reach
Today we have 3 tech labs.
Presence in these key technology
clusters gives an early view of
emerging technologies,
companies and start-ups.
San Francisco is today’s Innovation Capital of
the World and our tech lab will facilitate links
with innovative research start-ups, venture
capitalists and global technology leaders. This
office provides new opportunities at the
forefront of healthcare digital innovation and
major breakthroughs in enterprise technology.
Cambridge is the most dynamic scientific
business cluster in the world. We are
surrounded by 19 science parks with over
1,500 high-tech companies, world class
academic institutions at the bleeding
edge of scientific research and key
research hospitals like Addenbrookes.
Shanghai is emerging as the top city for
tech innovation internationally and is
predicted to become the global technology
centre within 4 years. Our tech lab is able
to tap into novel scientific research &
development, advanced engineering,
health nanotechnology and robotics.
5. UK Tech Lab
By having a highly technically skilled and business-savvy team of hands-on experts in our
Labs, we quickly assess new technologies and platforms in real-world settings, prioritising
on relevant business problems with strategic potential
5
It’s key for us to nurture the next
generation of IT leaders so we
work closely with new graduates
and local apprentice schemes.
Eddie Wu Izzy Derrig Charlotte Lorains Stefano Elia Sandra Giuliani Josh Mesout
IT Graduate IT Apprentice IT Apprentice IT Graduate IT Graduate IT Graduate
6. Technology Incubation
We’re designed to research customer problems, understand key insight, scout relevant
technology and then validate quickly with rapid prototypes and proof-of-concepts
6
We try to embrace Design Thinking and testing concepts early through Lean Start-up principles and
then iteratively exploring them further using Agile Development cycles and bi-weekly sprints.
7. Big Data in Pharma
7
AstraZeneca has been tackling big data challenges for over a decade.
Volume
SCALE OF DATA
Variety
DIFFERENT FORMS OF
DATA
Velocity
ANALYSIS OF STREAMING
DATA
Veracity
UNCERTAINTY
OF DATA
Next Generation Sequencing
Whole body imaging
Tissue Microarrays
Sales force optimisation
Clinical trial statistical analytics
High Throughput Screening
Toxicogenomics
Open Innovation Approaches
PowerPoint/Excel content
Structured databases
Predictive Chemistry Modelling
HR employee retention
Real-time news sentiment
experimental data capture
Wearable sensor information
Log analytics in Operations
“Big Data” is never solved by a single technology. It’s not a giant data mart or data model to solve it all
The best way to tackle big data challenges is to try lots of small proof-of-concepts to work out
which actually make a real difference to your business.
8. Unstructured Data
8
We have silos of unstructured content, both inside the company and in the cloud.
We initially focused on developing a big-data engine for unstructured R&D content for scientists.
Admin
Access
Data Sources
Data
Mappings
Configure
Permissions
Tag
Content
Sinequa
Index
Applications
Web Service
9. R&D Search
9
In 100 days we implemented Sinequa real-time search engine within R&D for
10k users. Covering all scientific information and core internal repositories.
Today, we have over
180 million documents,
searchable sub-second
with key scientific
vocabularies (SciBite)
automatically tagged
and findable.
Users can create alerts
to their favourite topics
but also find every
document relating to a
drug (and every
synonym automatically)
10. R&D Intelligence
10
We built R&D Intelligence to find things you don’t know about ! This computes sentence
level co-currence between any two entities to instantly spot new connections
Using this approach,
you can gain insight
and trends across the
entire data corpus, yet
provide visibility and
access to only the
material that you are
entitled to.
11. Enterprise Search
11
With 12 months experience of Sinequa, we developed a new global Enterprise
Search in 8 weeks and launched as part of the new portal, Nucleus. This now
has >60k users with multi-language support.
Search enables us to socialise key
findings from news & documents,
but also chatter, applications,
people and scientific tags –
helping to connect people
together.
Already in pre-production,
indexation of all of our cloud
repositories (Box, Sharepoint,
Veeva etc)
12. Big Data Engine
12
We had developed a big data engine that powers multiple
business applications, not just enterprise search. This
swiss-army knife for search let’s us tackle many problems
R&D News AlertsR&D ChemSearch
Find Partners
Mobile Apps Competitive Intel
Medical Affairs
13. Mobility: Find People
13
We leverage Sinequa as an MBAAS layer for multiple mobile apps
Find people based on all their information such as name, location, job title, biography and skills.
Call or email the person you’ve found at the click of the mouse or a touch of the screen.
See where a person fits in AstraZeneca’s organisation, who their manager is and if they have any direct reports.
14. Mobility: Approvals App
14
Imagine 1 mobile app that shows all of your employee requests and lets you
approve them easily, compliantly and anywhere in the world.
14
simple mobile app for approving
travel, expense, invoice and
procurement requests on the move
senior executives currently testing
beta version
systems across 2 core platforms
integrated today Designed to
handle any approvals, any system
Others now in planning for 2016
Users provided feedback that led to
rapid redevelopments to the
interface. Simplified the information
shown, however still compliant
Currently scaling up for potential roll-out for key users
1
50
10
15. Usability
15
Even with great feedback, wanted to continually improve usability. In 2015 we
helped establish a new UX competency center in AZIT.
15
Enterprise Search
“it finally feels that
you are being
listened to”
“Impressive! An AZ
search engine that
brings back documents
that are useful.”
“Wow ,
news on my
mobile!”
“I see how this
can help make us
a $50B company“
"In all my years, I have
never been able to find the
data until now"
“Can we use
this info in
our app?”
“This is a game changer”
“Can we use this
search for our system”
16. Watching Real Users
16
Helped us implement 40 quick fixes within 5 days but major overhaul to the UI.
Led to improvements in search and usability with 3 bespoke search apps.
16
Developed a schedule of UX lab tests for >10 major AZIT platforms to assess
usability and determine potential user-centered improvements for future releases.
Before After
17. Elastic Cloud Performance
17
Sinequa now handles ~2500 queries per min. Designed from the start to
leverage cloud elastic scaling capabilities for responsive performance for 10
users or 70,000 users.
17
Happy UsersDockerElastic Beanstalk
By leveraging container technology, we can spin up new services quickly. It’s very
cost-effective and enables new approaches to be easily included in your workflows
Automated video transcription
high precision text analytics
18. In-Video Search !
18
The Azure Media Services cloud platform is able to ingest audio and video files
and automatically generates a >90% accurate transcription (English only)
18
By combining Amazon, Docker and Microsoft Azure, we were able to go from first
prototype to global implementation in 7 weeks in our enterprise search platform.
19. Mobile + Video + Usability
19
Developed Proact – mobile app to capture the usability
and tolerability of our drugs in clinical trials through
patient video diaries.
19
• Built using UCD principles with multiple points of
view (patient, doctor, analysts, directors)
• From original prototype to pilot in 90 days
• Enterprise grade, HIPAA compliant, Secure.
• Available on the Google App Store
https://play.google.com/store/apps/details?id=com.astrazenaca.proact
20. Data & Log Analytics
2020
Even unstructured data engines generate structured data – but big data
opportunities live monitoring of logs from Sharepoint, Ping and Box.
21. Predictive Modelling
21
With larger, complex datasets, machine-learning techniques can accelerate and
increase accuracy of decision making, improving our productivity
Optimised models created using cloud elastic approaches offer another big data engine
opportunity across the enterprise:- salesforce optimisation, employee retention, resource
allocation, patient responder prediction, manufacturing/supply chain improvements.
22. The Next Big Data Engine ?
2222
We are really
excited to hear
about new,
emerging
technologies
that could be
the next big
data platform
with enterprise
applicability
across
AstraZeneca.
In true lean-start-up fashion, we are testing something new. If you are a start-up with an
emerging technology that seems relevant, please pitch IT at http://pitchIT.astrazeneca.com